Can anyone shed light on Famicom Disk serial numbers?

[DefaultGen | 5,401]

I'm looking into the actual serial numbers on the disks. I didn't find anything with a quick search. I don't want to reinvent the wheel if someone already knows what these mean. Zelda has a lot of them, which makes sense since Zelda probably was in print for a very long time so that's a good place to start. At a glance I see Zeldas with serials starting with H, I, J, K, and L. The second letter seems to be I, J, or K from what I've seen. Looking at a late release like Time Twist, the serial isn't too far off from the average Zelda, so it might not be related to date.

I'm hoping these can be used to help authenticate rare games or piece together correct inserts with certain disk variants, but I'll need a lot more information since I'm at square zero here. All the Charumera Zeldas I see at a glance are one digit off from each other so I doubt they're completely random.

Some random Zelda 1 disk codes in case that helps anyone figure it out:

K15J04
I296I04
J145I24
J316I22
J026I03
K215J29
K296K10
K155J11
K086J16
K135J11 (Charumera)
K145J11 (Charumera)
K195J11 (Charumera)
L185K38

 

Time Twist, a game that wasn't released until 1991:

Charumera Zelda:

Edited June 22, 2021 by DefaultGen

[Hammerfestus | 3,833]

Thought maybe from first 2 pics that your second letter is like country of manufacture?  But third doesn’t support that.

[0xDEAFC0DE | 1,277]

Huh, never noticed those before. I think I figured it out (although there still is some weirdness). First, I looked at several more games. I started with Janken Disk Jō (the last FDS game), but I quickly realized that the Japanese version of the wikipedia List of FDS games notates which games were exclusively packaged for sale and not offered Disk Writers. So, I found a bunch of codes for those games as well. Here's the list.

Janken Disk Jō: 12/22/92
K276K10
J286I22
H216G07
H276K10
I206I03?

Hikari Genji: Roller Panic: 3/20/89
D198C10
E068C10

Chitei Tairiku Ordola: 3/27/87
B047L15
C187B23?
K286K10?

Super Boy Allan: 3/27/87
A197D15

Kieta Princess: 12/20/1986
L066K17
L046K17
L056K17
K256J22
L046K17
L086K17
L046K17
L046K17
L056K17

So, a couple things off the bat. First, the letters do use the full range of A-L. That's the first 12 letters which obviously suggest a month code. The first two numerical digits do not seem to go over 31 suggesting that they represent the day. Finally, the third numerical digit roughly matches up with the last digit of the year. So, I think the date code format is MDDY???. Looking at Zelda and the four exclusively packaged FDS games the dates mostly line up although they do seem a bit early. I'm wondering if that is when the disks were manufactured and not necessarily when the game was written to the disk. That would also explain why Janken Disk Jō dates don't match up at all.

I'm also not sure on the last three characters. The second letter also seems to range from A-L and the last two numbers don't go above 31 (except for one you posted which you might have misread). So maybe that's another date that indicates something else? But I've not seen any indication that they were printed at separate times, so IDK.

[RetroGE | 21]

17 hours ago, Hammerfestus said:

Thought maybe from first 2 pics that your second letter is like country of manufacture?  But third doesn’t support that.

I was thinking the same.  Or maybe city? 

[RetroGE | 21]

5 hours ago, 0xDEAFC0DE said:

Huh, never noticed those before. I think I figured it out (although there still is some weirdness). First, I looked at several more games. I started with Janken Disk Jō (the last FDS game), but I quickly realized that the Japanese version of the wikipedia List of FDS games notates which games were exclusively packaged for sale and not offered Disk Writers. So, I found a bunch of codes for those games as well. Here's the list.

Janken Disk Jō: 12/22/92
K276K10
J286I22
H216G07
H276K10
I206I03?

Hikari Genji: Roller Panic: 3/20/89
D198C10
E068C10

Chitei Tairiku Ordola: 3/27/87
B047L15
C187B23?
K286K10?

Super Boy Allan: 3/27/87
A197D15

Kieta Princess: 12/20/1986
L066K17
L046K17
L056K17
K256J22
L046K17
L086K17
L046K17
L046K17
L056K17

So, a couple things off the bat. First, the letters do use the full range of A-L. That's the first 12 letters which obviously suggest a month code. The first two numerical digits do not seem to go over 31 suggesting that they represent the day. Finally, the third numerical digit roughly matches up with the last digit of the year. So, I think the date code format is MDDY???. Looking at Zelda and the four exclusively packaged FDS games the dates mostly line up although they do seem a bit early. I'm wondering if that is when the disks were manufactured and not necessarily when the game was written to the disk. That would also explain why Janken Disk Jō dates don't match up at all.

I'm also not sure on the last three characters. The second letter also seems to range from A-L and the last two numbers don't go above 31 (except for one you posted which you might have misread). So maybe that's another date that indicates something else? But I've not seen any indication that they were printed at separate times, so IDK.

Wow!  Great sleuthing! I’m sitting on a couple hundred disks and some free time.  I’ll dig in on my end.  As a side note my 2 copies of Kaettekita Mario have codes of FD47E09 and J106I22.  This was a disk writer exclusive so it would make sense those codes could be all over the place

[DefaultGen | 5,401]

Yeah the disk writer games would make things tricky, but good job! I gotta make a little converter for all these so I don't have to keep counting on my fingers.

[DefaultGen | 5,401]

I need to stare and compare more codes to figure out the date code on the end. More of them seem to match (just look at how many Kieta Princesses are K17s).

The digits can be a little offset sometimes, not sure if that can mean anything (like it was stamped twice?). The later date can be either before or after the first date. A mystery

[Sho | 25]

Hello! 

A friend brought this thread to my attention and this is a very interesting find indeed.  And I have to say you guys are DEFINITELY onto something here concerning the stamped date codes. 

I'll preface this with the idea that you must first exclude any and all Disk Writer games for the time being (a few games might be exempt to this rule).  DW games will in almost all cases be re-written onto other games, therefore overwriting the original write date and negating the stamped code on these disks.  Prototypes and sample disks (white disks) must also be excluded as they were constantly written and rewritten for dev work.

With that said, there is a FDS program that a gentleman known as Chris Covell created in 2001 (and updated recently for 2021) that will actually read the header contents of a Famicom Disk.  This program is known as the FDS Disk Lister and can be copied onto a disk and be used with a real Disk System.  This program will reveal the disk write date as well as another date that can be speculated on as the "retail" or "copyright" date.

Using this program in tandem with the methodology that 0xDEAFC0DE arrived at I did a few examples:

Given the code on this Akumajou Densetsu, the write date should be September 16, 1986.  This is what the Disk Lister program revealed:

 

Written in September 24, 1986!  The date is off of course, but this might simply mean that the game was copied onto the disk a few days after the date was stamped.  For games that were created en-masse, it would make a lot of sense to have them prepped, then copied a few days later.  What is also interesting about this is that this copy date also predates what was the assumed release date of this game by a few days, September 26, 1986.

For added kicks, in this example you can see a Game Version code as well as a Disk Version (Debug) code.  These codes are just as you might think they are, they are revision codes, similar to the REV-A stuff you'll see on various NES carts.  Not all games have more than one revision.  However some games have many.  Akumajou Densetsu has 3 revisions! (The above is Ver 1.0, or REV0)  As you can imagine each newer revision comes with various bug fixes and minor improvements and so on.

Moving onto the next example:

Given the stamped code the date should be December 01, 1986.  Here's what the program revealed:

 

December 04, 1986.  Another home run on that one.  However, this also reveals another facet of the stamped/copy dates -- they only really pertain to when the disk was written.  Zelda no Densetsu was originally released in February 21, 1986, not December.  The Game/Disk Version denotes this as Ver 1.1, or REV1.

And yet another disk, on another year:

Stamped code gives a date of January 19, 1987.  Here is what was revealed:

 

January 27, 1987.  You guys really nailed down the first few numbers.  As long as you respect it, you can look and get an approximate date when the game was pressed at a glance.  You'll also need to respect holidays as well (there was an instance with my copy of Legend of Zelda II -- the disk was pressed on December 25, 1986, but the data wasn't written until January 12, 1987).  Of course, the Famicom Disk System is known for being the Wild Wild West and it's not always the case.  Different games get written and re-written all of the time, and if you want to be sure that a game hasn't been touched, you'll want to aim more for complete in box copies. 

With that said, I can't for the life of me figure out the trailing codes after the date.  It's very possible it might have to do with where it was stamped, be it in another machine, or in another location entirely.  Or it might denote something else that is internal to Nintendo.

Hopefully the program helps you guys discover more with these disks, it is fun looking at these stamped codes and I'll be paying attention to them from here on out.

Edited June 18, 2021 by Sho

[Deadeye | 1,596]

Great research and info @Sho.  Thanks!

[0xDEAFC0DE | 1,277]

Awesome work @Sho! Do you know of any games that have a different disk write date and "copyright" date? I'd be curious to see which date matches up with the stamped date.

[DefaultGen | 5,401]

God I love VGS

[noiseredux | 232]

47 minutes ago, DefaultGen said:

God I love VGS

right? This thread is fascinating.

[Sho | 25]

Welcome gentleman, my pleasure! 

9 minutes ago, 0xDEAFC0DE said:

Awesome work @Sho! Do you know of any games that have a different disk write date and "copyright" date? I'd be curious to see which date matches up with the stamped date.

I actually do have a few of these disks, and they tend to all be oddballs.  I call them "oddballs" because they are all written or re-written games by various methods over the last 30 years (official Disk Writer kiosks, official "internal" rewriting tools, Dubbing tape, Tonkachi Editor, Game Hacker, FDSStick, FDSLoader and so on). 

However, when the official Disk Writer is concerned, it seems to continue the trend and expose the "copyright" date as original and "initial" writing date for the disk itself.  The first example:

 

This is the game that was originally written as... something, in late October, Early November 1986.  The label is an original Section Z label (original release date May 25, 1986).  Note that you can still see the residue where the original label had been, whatever that game may have been.  Here's what the Disk Lister exposed:

 

Pro Golfer Saru... written in May 07, 1988.  Note that not only is the written and "copyright" dates are now different, but the "copyright" date still follows the stamp date of the disk, November 11, 1986.  This also exposes the fact that not only did the game get copied once from it's original game, but twice, if not more between 1986 and 1988.  They simply neglected to stick on a new label at some point.

Next example:

 

A game that was written for the Disk Writer only game Galaxian with it's label attached hilariously upside down (Galaga is on Side B, also with it's label upside down, hehe).  In the picture the stamped code is faded but it reads A287L15, which should date January 28, 1987.  Here's the Listing:

 

Here many things prove that it was taken to an official Disk Writer for copying.  The "copyright" date follows the "initial" copying date, if not a month late from the disk's stamped date, February 26, 1987.  The written date is September 22, 1990, two days after Galaxian's official DW release date to Japan.  One could imagine the labels were hastily placed onto the disk in error as they were handling a line of people eager to play this at the time new release to disk.  Alternatively, I like to believe that the person requested to have the label placed upside down, haha.  I checked a few more official Disk Writer releases and they all seem to follow this path as viewed on the Lister.

Unfortunately, is this where the zaniness of FDS piracy/bootlegging/future copying methods comes in.  Moving beyond the official Disk Writer kiosk, everything gets thrown out the window concerning the stamped/"copyright"/written dates.  The first example:

 

This is the FDS Disk Listing program itself that I have been using.  When I was told of this very thread, I took out the time to copy the newest revision of the program back onto this disk (Using FDSStick and an old revision FDS Drive), as I had been lazily using v1.8 for the last few months.  The disk's stamped date denotes September 01, 1986.  Keep in mind I copied this program today, June 18, 2021.  Here's the meta that was revealed:

 

The "copyright" date as well as the written dates are both completely off!  Actually, the written date makes a bit of sense... it's the date that Chris Covell pushed his newest revision of the program, as seen in the title above.  I then remembered:  In the NESDev forum, I actually asked Chris about the "copyright" date a few months ago.  Here was the exchange:

Me: "I see! What about the CopyRight date? Should that number be consistent amongst different revisions of the same game? If not what could that mean? From there, I assume that the "Written" date can be overwritten and changed if it was taken to a local kiosk as well, yes? Assuming the kiosk itself must have had an internal clock of some kind.

And finally, what of the the disks that have completely blanked out dates?"

Chris: "Your guess is as good as mine for all the above. Different versions of games, for example, Zelda, have different copyright/manufactured dates. Rewritten dates would probably be set by the machines in Nintendo's factories, or by the Disk Writer kiosks' running clocks.

Plenty of games have invalid date codes -- they're supposed to be stored in BCD, so any hex bytes $a-$f renders them invalid, along with a date of all zeroes.

The header contains some bytes useful to FDS units, but a lot of them appear only to be used for version tracking. Plenty of them are not even filled in. For example, the prototype of Air Fortress on the FDS has some text filling up the FDS header!
see:..."

https://forums.nesdev.com/viewtopic.php?f=2&t=22137&start=15#p270312

Hyperlink directly to that post if you would like to read it.  Unfortunately, the chatter kinda died down right after that post.  This thread is now it's spiritual successor, hehe.   If you scroll a bit up some, we also discussed how FDSStick copying onto disks also respects the original disk headers of games written at official kiosks.  What this will mean is that if you say, for example found a ROM of an FDS game online to write back to a dead disk, it will also write back it's written and "copyright" games for that ROM that was dumped.  What this also should mean is that Chris developed his program, changing the "copyright" date to 2002 himself for whatever reason.

In that exchange I asked about "completely blanked out dates".  This refers to the next example:

 

 

This is a prototype/sample disk of Ginga Denshou: Galaxy Odyssey.  The disks stamped date should denote November 05, 1987.  Here's what the program revealed:

 

February 05, 1987 for the "copyright" date.  Also the written date is completely blanked out.  This could be seen as somewhat logical, seeing as it's a sample/prototype disk and wasn't in production.  However... the problem with this is that the game Ginga Denshou was already released in November 11, 1986.  Not only this, but 2 revisions of the game are known, REV0 and REV1,  If you look at the Game Version in this example it denotes a REV2.  It's likely possible that a REV2 exists in the wild and hasn't been properly catalogued, and this was the test disk to check for bugs before they pushed the revision.  Alternatively, it might also mean that they canned the revision altogether. This game in particular was a game that came in a complete set, with a hint booklet, audio tape (for storyline), lyrics booklet and so on:

Stuff like this isn't always the case with protos, but you should be wary of the stamped dates on these guys to be sure if you acquire them, as they more often than not will not follow the rules. 

Another example.  If you remember above I said in my prior post that Akumajou Dracula* (Not "Densetsu") has 3 revisions:

 

A stamped code that denotes May 09, 1986.  Here is what the Lister revealed:

 

A "copyright" date of July 07, 1986, but a write date of January 25, 1987.  Paying attention to the Game Version, you will also see that it reveals this game as a REV2 (or v1.2) of this game (the prior example was REV0).  What this also shows us is that this disk was taken to a Disk Writer for copying -- recall that Akumajou Dracula was released in September 1986, while it's disk and it's prior contents was created in July.  There is an old timer Japanese Famicom Disk System enthusiast who has been tracking game revisions for years.  He hypothesized that all REV2 copies of Akumajou Dracula are all Disk Writer releases.  In this instance, it follows up with his on-going theory.

If you would like to see this revision listing that I'm referring to, please click on this link: https://pony.velvet.jp/fcdisk/fdsstdylst_e.html

We could possibly do this all day and night btw!  There are so many zany things that can be gleamed from using the stamped codes and this listing program in tandem. 

[Sumez | 2,939]

So did you guys figure it out?

I have a crap ton of FDS games I can dig out if it helps anything

[DefaultGen | 5,401]

Looks like I need a FDSStick to investigate more  So one anomaly which isn't a typo is K15J04. @Sho do you have any disks with 6 character codes? My only thought is early disks didn't make sure the date was always 2 digits (e.g. it's 1 instead of 01). From at least 1986 on at least, days 1-9 are 2 digits.

So K15J04 could really be K015J04, which would put it right in line with other 1985 date codes (November 1). The earliest Zelda I've seen so far is J145I24 (October 14) and there are November 15 and 21 dates too.

Edited June 22, 2021 by DefaultGen

[Sho | 25]

On 6/20/2021 at 5:36 AM, Sumez said:

So did you guys figure it out?

I have a crap ton of FDS games I can dig out if it helps anything

It looks like for the most part 0xDEAFC0DE figured it out!

1st Digit:

A = January
B = February
C = March
D = April
E = May
F = June
G = July
H = August
I = September
J = October
K = November
L = December

2nd & 3rd Digit:

01 - 31 for the day

4th Digit:

6 = 1986

7 = 1987

8 = 1988

9 = 1989

We'll have to see some examples of 1990 being stamped onto a disk, and "1985" would need to be tested too.  As far as the last 3 digits, it's still unknown, but it could be speculated on being either the location of the printing, the machine number that printed it or a combination of both.

 

Edited June 23, 2021 by Sho

[Sho | 25]

12 hours ago, DefaultGen said:

Looks like I need a FDSStick to investigate more  So one anomaly which isn't a typo is K15J04. @Sho do you have any disks with 6 character codes? My only thought is early disks didn't make sure the date was always 2 digits (e.g. it's 1 instead of 01). From at least 1986 on at least, days 1-9 are 2 digits.

So K15J04 could really be K015J04, which would put it right in line with other 1985 date codes (November 1). The earliest Zelda I've seen so far is J145I24 (October 14) and there are November 15 and 21 dates too.

This is very possible that it is a mistake and 0 is missing missing after the K!!!  If it's indeed a November date, you could imagine it being part of an enmasse printing 2 months prior to the Zelda no Densetsu's release.  If we want to use this line of thinking, these games should be checked for possible errors:

Baseball

Golf

Mahjong

Soccer

Super Mario Bros.

Tennis

Zelda no Densetsu

All of these games were released in the same time frame and was likely subjected to a very large initial print run, likely in late 1985 too.  I'll check my games tomorrow!  I have a few copies of some of these games.

 

And I totally recommend getting an FDSStick!  it's a very versatile device and comes in handy when you need to spoof a disk you don't have, have a small FDS collection that's portable, restoring data on a dead disk, amongst other things like drive calibration and more.  I'm waiting to get two more for testing Game Doctor disks myself.  However it's still sold out!  Check here occasionally:

http://3dscapture.com/fdsstick/

Edited June 23, 2021 by Sho

[DefaultGen | 5,401]

Not the biggest update in the world but I have a K095J11 Zelda, so if the missing zero theory is true then it was fixed about a week later. Could just be a one time thing where they forgot the zero on the K15J04 ones though. I'm waiting on a FDSStick restock!

[Sho | 25]

On 6/28/2021 at 2:15 PM, DefaultGen said:

Not the biggest update in the world but I have a K095J11 Zelda, so if the missing zero theory is true then it was fixed about a week later. Could just be a one time thing where they forgot the zero on the K15J04 ones though. I'm waiting on a FDSStick restock!

Heya!  A bit late in reading this (OK, hella late!) But I'm here to tell ya the FDSStick has been restocked!

http://3dscapture.com/fdsstick/

[AdamW | 713]

I ran across another 'anomalous' Zelda:

https://www.catawiki.com/en/l/34555665-nintendo-famicom-disk-system-2-zelda-games-both-complete-with-unused-stickers-2-in-original-box

the code is K75J11. So if our theory is true they added the 0 later on November 7th, any time on November 8th, or sometime on November 9th.

[Sho | 25]

On 10/7/2021 at 12:33 PM, AdamW said:

I ran across another 'anomalous' Zelda:

https://www.catawiki.com/en/l/34555665-nintendo-famicom-disk-system-2-zelda-games-both-complete-with-unused-stickers-2-in-original-box

the code is K75J11. So if our theory is true they added the 0 later on November 7th, any time on November 8th, or sometime on November 9th.

That sounds about right!  It does not deviate too far off the mark unless it was rewriten or a DW variant